|Home | About | Journals | Submit | Contact Us | Français|
Genetic rearrangement of TMPRSS2 regulatory sequences and coding sequences of the ERG gene has been detected in nearly half of prostate cancers. Quantitative assays to detect such TMPRSS2-ERG gene fusion have been limited to real time PCR techniques that rely on reverse transcriptase-based amplification. We sought to develop a novel assay that uses branched DNA (bDNA) technology to measure TMPRSS2-ERG fusion.
Branched DNA probes were designed to detect TMPRSS2-ERG gene fusion in prostate cancer cell lines. Non-quantitative, nested reverse transcription (RT)-PCR and fluorescence in situ hybridization (FISH) were used to ascertain TMPRSS2-ERG gene fusion status in prostate tissues.
The branched DNA assay detected TMPRSS2-ERG gene fusion from less than 200 picogram of prostate cancer RNA, whereas more than 600 picogram of RNA was required for fusion gene detection by one step real time RT-PCR. In evaluation of clinical prostatectomy specimens, the branched DNA assay showed concordant detectable fusion signal in all 9 clinical samples that had fusion detected by nested RT-PCR or FISH. Moreover, branched DNA detected gene fusion in 2 of 16 prostate cancer tissue specimens that was not detected by FISH nor nested RT-PCR.
Our findings demonstrate a branched DNA assay that is effective for detection of TMPRSS2-ERG gene fusion in prostate cancer clinical specimens, thus providing an alternative method to ascertain the TMPRSS2-ERG gene fusion in human prostate cancer tissue.
Tomlins et al. have reported a recurrent fusion of the androgen-regulated gene TMPRSS2 to the ETS transcription factors ERG, ETV1, and ETV4 in prostate cancer.1,2 Subsequently, multiple studies have confirmed the presence of TMPRSS2-ETS gene fusions, especially TMPRSS2-ERG as the most common, which is present in approximately 50% of prostate cancers.3-6 ERG has been previously identified as an overexpressed gene in prostate cancer by real time PCR using primers from 3′ portion of the ERG gene, distal to the region involved by the TMPRSS2 fusion with demonstrated increase of ERG mRNA expression in 62% of prostate cancers7. Wang et al. further correlated the isoforms of TMPRSS2-ERG fusions and expression levels with clinical outcome in prostate cancers from patients undergoing radical prostatectomy.5 TMPRSS2-ERG fusion transcripts thus also have the potential to serve as a biomarker for a more aggressive clinical course of prostate cancer. Moreover, TMPRSS2-ERG fusion transcripts can be detected in urine obtained from patients with PCa after prostatic massage, suggesting possible avenues for clinical use of TMPRSS2-ERG fusion detection.8
Assays for detecting TMPRSS2-ERG fusion have been limited to those based on RT-PCR or FISH. RT-PCR requires the presence of a stable, full-length transcript that can be difficult to retain in routine clinical processing, whereas FISH requires subspecialty molecular pathology expertise that is not uniformly available. In contrast, branched DNA techniques for detecting expressed transcripts use in-solution hybridization followed by cooperative hybridization between a target RNA and a target-specific probe set in an assay platform that requires neither reverse transcription and target amplification (as required by RT-PCR) nor esoteric expertise (as required to interpret FISH in situ).9,10 Furthermore, bDNA is apparently less dependent on RNA quality than RT-PCR”. 11 We therefore developed a novel branched DNA probe design to measure the TMPRSS2-ERG fusion mRNA expression in human prostate cancer specimens.
Vcap cells were purchased from American Type Culture collection (ATCC, Manassas, VA).
Radical prostatectomy tissue samples were obtained from the IRB approved Hershey Foundation Prostate Cancer Serum and Tumor Bank at Beth Israel Deaconess Medical Center (BIDMC). Morphologic diagnosis was performed by a pathologist on hematoxylin and eosin (H&E) stained slides made from both sides of the corresponding (OCT)-embedded prostatectomy tissue blocks using an optimal cutting temperature. OCT blocks containing more than 30% of prostate cancerous tissue (with Gleason Score of 6 or 7) were selected for RNA purification.
A biopsy punch (Miltex Inc., York, PA) was used to select the prostate cancer tissues from the OCT sample blocks. Benign or prostate cancer tissues were homogenized using a TissueLyser (Qiagen, Valencia, CA) at 28 Hz for 5 min. Total RNA was isolated using TRIzol reagent (Invitrogen Corporation, Calsbad, CA) according to the manufacturer's instructions. RNA integrity was verified using an Agilent 2100 Bioanalyzer. Samples with high RNA integrity number (RIN) > 7 were selected for analysis.
Primers for TMPRSS2-ERG fusion detection by nested RT-PCR were TMPRSS2-1F 5′-CGCGAGCTAAGCAGGAGGCG-3′, ERG-541R 5′-TCATGTTTGGGGGTGGCATGTG-3′, TMPRSS2-20F 5′-GGAGGCGGAGGCGGAGGG-3′and ERG-450R 5′- TTGGCCACACTGCATTCATC AGGA-3′. 1μl of cDNA template was amplified in a final volume of 50 μl using Platinum Tag DNA Polymerase (Invitrogen Corporation, Calsbad, CA) according to the manufacture's instructions. 2 μl of first round PCR product was used as template for nested PCR (primers: TMPRSS2-20F and ERG-450R). SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA) was used for one-step real-time RT-PCR analysis on Applied Biosystems 7900HT Prism instrument. Primers for TMPRSS2-ERG fusion detection by real time PCR were TMPRSS2-1F 5′-CGCGAGCTAAGCAGGAGGCG-3′, ERG-541R 5′- TCATGTTTGGGGGTGGCATGTG-3′. 50 μl of cDNA was synthesized from 10 ng total RNA and 1 μl cDNA is analyzed by real time PCR (two-step real-time PCR). Samples are performed in triplicate. Relative quantification analysis was used. Expression values of each gene were normalized to the expression of GAPDH of a given sample. The 2−ΔΔCt method (available online from Applied Biosystem) was used to calculate relative expression of each gene.
FISH analysis of the specimens was performed as previously described by Perner and Mosquera. 1,3,12 There were two differentially labeled probes designed to span the telomeric and centrometric neighboring regions of the ERG locus. A nucleus without ERG rearrangement gives a yellow signal (juxtaposed red and green signals). TMPRSS2-ERG fusion through insertion results in a single red and green signal for the rearranged ERG allele and a yellow non-rearranged allele in each nucleus. TMPRSS2-ERG fusion through deletion results in a loss of the green telomeric signal with a remaining red signal for the rearranged allele and a yellow signal for the non-rearranged allele.
Experiments addressing optimal capture extender (CE), blocker (BL), and label extender (LE) concentrations indicate that these are 25 fmol, 50 fmol, and 100 fmol respectively across a broad range of bDNA assays. 10 ng of RNA were mixed with the working probe set and lysis mixture according to the QuantiGene 2.0 Reagent System manual (Panomics, Fremont, CA). 100 μl/well were then dispensed into the capture plate. Hybridization was performed overnight (16-20 hr) at 55°C and the plate was then processed following the protocol. the capture plate was read on Victor3 1420 multilabel counter (Perkin Elmer, Inc. Waltham, MA).
Detection sensitivities of real-time RT-PCR and bDNA technologies spiked in total RNA from Vcap cells were each determined by ANOVA modeling with pairwise comparisons of each dilution level versus the control. The TMPRSS2-ERG fusion gene expression measured by real-time RT-PCR and bDNA were compared between cancer and controls using Wilcoxon rank sum test. In addition, signal strength of each cancerous sample was compared to the 95% confidence limits of the mean of benign samples, and signals that exceeded control 95% confident limits were designated as indicating significant expression.
Branched DNA (bDNA) probe sets for the fusion gene were designed to capture the 5′ portion of the TMPRSS2 gene and exons 5-6 of the ERG gene (Figure 1). Sequence information regarding capture extender and label extender are shown in Figure 1. To compare the detection sensitivity of the TMPRSS2-ERG fusion by real-time RT-PCR and bDNA technologies, RNA from Vcap cells, known to express the TMPRSS2-ERG fusion, was combined with total RNA from benign human prostate to a total of 10 ng RNA. First, we tested the sensitivity of the bDNA technology by using an ANOVA model (Figure 2A). TMPRSS2-ERG fusion was detected by significant luminescence (p< 0.05), at 0.156 ng or higher amounts of VCAP RNA (Figure 2A). The sensitivity of real time RT-PCR for detection of the TMPRSS2-ERG fusion is shown in Figure 2B and was limited to 0.625 ng or higher amounts of Vcap RNA (p<0.05).
We next sought to determine if the increase in assay sensitivity, in terms of RNA amount required to detect fusion, would translate to improving the ability to detect TMPRSS2-ERG fusion in 20 clinical prostatectomy specimens. We first evaluated 4 benign and 16 cancerous prostatectomy samples by nested RT-PCR to determine their fusion status; a subset of the cancers' fusion status was ascertained by FISH detection of translocation or deletion (Table 1; insufficient sample tissue blocks precluded conclusive FISH assay in 5 cases). By nested RT-PCR, 9 of the 16 cancerous samples were fusion-positive (all 4 benign samples were fusion negative; Table 1). In 2 cases (cases 04-33LPB1-T3 and 04-33LPB4-T3) evaluable by FISH, FISH did not detect fusion that was detected by nested RT-PCR.
After ascertainment of fusion status by the non-quantitative nested RT-PCR technique, we next turned to a comparison of the sensitivity of real time PCR and branched DNA assays in detecting TMPRSS2-ERG gene fusion. Real-time RT-PCR failed to detect TMPRSS2-ERG gene fusion in 1/9 cancer cases that had fusion detectable by nested RT-PCR (case 04-34RB1-T1; Table 1). In contrast, the branched DNA assay detected TMPRSS2-ERG gene fusion in 9/9 cases that had been detected by nested RT-PCR and also detected the fusion in 2/16 cases that had not been detected by nested RT-PCR at 95% CL (cases 04-26L2 and 05-04RA1; Table 1).
Identifying specific cancer target genes is of paramount importance for early detection of the disease and the diagnosis of disease progression. To date, two common chromosomal aberrations have been identified in prostate cancer; the androgen receptor (AR) gene at Xq12 and TMPRSS2-ERG at 21q.13 with TMPRSS2-ERG fusion gene being the most common genetic aberration present in about half of clinically localized prostate cancer tissues. 3-6,14
Reverse transcription-PCR (RT-PCR), nested RT-PCR, RT-PCR followed by Southern blot hybridization with a radio labeled probe and FISH have been used to analyze the prostate cancer specific fusion genes.1,4,5,15 However, the sensitivity to detect TMPRSS2-ERG expression level using standard methods has not rigorously been evaluated. Currently, real time RT-PCR is the principal method for quantitative measurement of TMPRSS2-ERG fusion based on expression levels. Here we employed a novel probe set design based on the bDNA technology to detect the expression of the TMPRSS2-ERG fusion gene. The branched DNA assay described herein detected fusion transcripts at levels as low as 0.16 ng RNA. This sensitivity translated to clinical samples in the branched DNA assay successfully detecting fusion in all 9 positive prostate cancer samples found to have fusion by other non-quantitative methods (nested RT-PCR or FISH). Moreover, the branched DNA technique was able to detect TMPRSS2-ERG fusion in cases for which the standard FISH, RT-PCR, and real time PCR assays were not sufficiently sensitive to detect fusion. Possible reasons why the bDNA assay may detect TMPRSS2-ERG gene rearrangements not detected by nested RT-PCR include: 1) a gene rearrangement may occur outside of the nested RT-PCR primer sets, whereas the bDNA extender sequences flank beyond the internal nested primer sets by 185 base pairs; 2) the number of fusion-positive cells in a tissue specimen may be below the detection limits of nested RT-PCR but still be detected by bDNA due to the favorable inter-assay variance characteristics of bDNA-based techniques that do not rely on target gene amplification; 3) there were no real fusions in the two samples detected only by the branched DNA technique (i.e., these were false positive results). Only with additional detailed molecular studies will the true sensitivity and false positive rate of the branched DNA technique be determined. Concurrent studies suggest that branched DNA techniques in general may be applicable to detect gene rearrangement directly from formalin fixed, paraffin embedded tissues without any need for distinct RNA isolation (extraction).11
The differential expression of the fusion genes may provide useful information in clinical practice. Although the exact mechanism of TMPRSS2-ERG fusion is yet to be unraveled, this gene fusion is considered to be an early event in prostate cancer development.16 Under physiological androgenic stimulation, the fusion gene may lead to overexpression of ERG regulated genes involved in cell growth and differentiation.17 There is a significant link between TMPRSS2-ERG and a distinct phenotype in prostate cancer such as blue-ring mucin, cribriform growth pattern, macronucleoli, intraductal tumor spread, and signet-ring cell features. 12 The fusion gene also occurs in a proportion (21%) of high-grade prostatic intraepithelial neoplasia (HGPIN) lesions, possibly preceding chromosome copy number changes in prostate carcinomas, and has been associated with invasion.16,18 In a pool of 26 patients who underwent surgery for clinically localized prostate cancer, patients with the fusion gene had a significant higher rate of recurrence (5-year recurrence rate of 79.5%) compared to patients without the fusion gene (five-year recurrence rate of 37.5%). 19 The TMPRSS2-ERG rearrangement has been implicated with prostate cancer clinical aggressiveness, but such an association has not been definitively confirmed5,20,21. The TMPRSS2-ERG fusion was also reported to be associated with lethal prostate cancer in a watchful waiting cohort (cumulative incidence ratio = 2.7, p < 0.01, 95% confidence interval = 1.3 – 5.8). 20 Furthermore, a study showed that TMPRSS2-ERG fusion was observed in androgen receptor (AR)-negative xenograft and in clinical prostate cancer specimens, while the fusion gene was not expressed. However, two other wild-type ETS family genes, ETV4 or FLI1, were over-expressed in these AR-negative tumor samples. 22
In addition to the probable role the TMPRSS2-ERG fusion plays in the pathogenesis of prostate cancer, it also has the potential to serve as a biomarker of prostate cancer with more aggressive behavior as compared to the fusion negative prostate cancers. In this scope, TMPRSS2-ERG fusion transcripts have been detected in urine obtained after prostatic massage.9 Through a combination of TMPRSS2-ERG and prostate cancer antigen 3 (PCA3) assay, detection of transcripts in urinary sediments could achieve a sensitivity of 73% in a prostate cancer positive biopsy group. Given the specificity of TMPRSS2-ERG for prostate cancer, the sensitivity of prostate cancer diagnosis could be significantly improved. A positive detection for TMPRSS2-ERG fusion could prevent repeat biopsies of those patients with elevated serum PSA levels and a history of negative biopsies. Two out of 30 men with a prostate cancer-negative biopsy (7%) had detectable TMPRSS2-ERG fusion transcripts in their urinary sediments.15 Whether the branched DNA technique for detecting TMPRSS2-ERG fusion can be applied to readily accessible clinical specimens such as urine awaits further study.
With TMPRSS2-ERG fusion gene being the most common genetic aberration in prostate cancer, TMPRSS2-ERG fusion defines a subtype of prostate cancer patients. The branched DNA method described herein is an effective method for detection of TMPRSS2-ERG fusion gene expression in prostate cancer tissue samples that allows fusion to be identified. This novel technique may help optimize detection of TMPRSS2-ERG gene fusion and thereby provides a potentially useful tool for prostate cancer molecular staging and detection.
This study was surpported by NIH-NCI Early Detection Research Network grant UO1-CA11391
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.